Background: Online longitudinal surveys may be subject to potential biases due to sample attrition. This study was designed to identify potential predictors of attrition using a longitudinal panel survey collected during the COVID-19 pandemic.

Methods: Three waves of data were collected online during the pandemic using Amazon MTurk. Variables included respondents’ demographics, medical history, socioeconomic status, COVID-19 experience, behavioral change, employment change, financial spending change, and health-related quality of life (HRQoL). Results were compared to US pre-pandemic norms. Measures that predicted attrition in the following wave were identified via logistic regression with stepwise selection.

Results: 1,467 out of 2,734 wave 1 respondents participated in wave 2 and, 964 out of 2,454 wave 2 respondents participated in wave 3. Younger age group, Hispanic origin (wave 1: p<0.001; wave 2: p=0.001), and higher self-rated survey difficulty (wave 1: p=0.002; wave 2: p<0.001) at waves 1 and 2 consistently predicted attrition in the following wave, respectively. COVID-19 experience, employment status, and limited physical activities were commonly observed factors contributed to attrition while specific measures change to reflect the most concerning matter at the time. In addition, mental health, average hours worked per day (p=0.004), and COVID-19 impact on work productivity (p<0.001) at wave 1 were found to be correlated with a higher attrition rate at wave 2. Support of social distancing (p=0.032), being Republican (p<0.001), and having just enough money to make ends meet (p=0.003) were some of the remaining key characteristics that predicted attrition at wave 3.

Conclusions: Attritions found in this longitudinal COVID-19 panel survey was not at random. Besides commonly identified demographic factors that contribute to panel attrition, COVID-19 presented new challenges to address sample biases by correlating attrition with additional factors in a constantly evolving environment.